> -----Original Message-----
> From: Margaret Wasserman [mailto:mrw@windriver.com]
> Sent: Monday, May 05, 2003 8:03 AM
> To: Brendan Kelly
> Cc: 'Andy Bierman'; 'Allen, Keith'; 'xmlconf@ops.ietf.org'
> Subject: RE: WSDL
>
>
> Hi Brendan,
>
> At 03:42 PM 4/16/2003 -0600, Brendan Kelly wrote:
> >Why can't we have both? I would like to see this working group to focus
> >on the application side (possible WSDL). Use the current draft from Rob
> >Enns as the basis for the configuration definition of the WSDL. Let CLI
> >be what it is.
>
> This would be fine, I suppose, if operators were happy with
> proprietary CLIs and screen scraping, but they are not.
>
I did not mean to suggest this working group deal with CLI. CLI is legacy and should stay that way.
> XMLCONF is a proposal to keep the advantages that have made
> CLI successful, while eliminating the serious shortcomings of
> CLI. It is most likely to be used where CLI is used today --
> at the interface between the NMS/EMS system and the network
> device.
>
I live and breath configuration management and completely understand the benefits of a device actually having an API.
> There is nothing about XMLCONF that precludes using a WSDL
> based enterprise management system on the application side.
> But, unless I'm misunderstanding what you mean by the
> "application side", it wouldn't make very much sense to
> standardize that in the IETF.
>
I am suggesting standardizing the WSDL definition. This definition could include things such as capabilities, services (routing protocols, QOS...) and their states. It would also define how a vendor would extend the standard definition.
> Margaret
>
Brendan